Citation

Vermeulen J, Willard S, Aguiar B, de Witte LP. Assist. Technol. 2015; 27(1): 18-23.

Affiliation

a School for Public Health and Primary Care (CAPHRI) , Maastricht University, Department of Health Services Research , Maastricht , The Netherlands.

Copyright

(Copyright © 2015, Rehabilitation Engineering and Assistive Technology Society of North America, Publisher Informa - Taylor and Francis Group)

DOI

10.1080/10400435.2014.949015

PMID

26132221

Abstract

The objective of this study was to evaluate the sensitivity and specificity of a smartphone-based fall detection application when different smartphone models are worn on a belt or in a trouser pocket. Eight healthy adults aged between 18 and 24 years old simulated 10 different types of true falls, 5 different types of falls with recovery, and 11 daily activities, five consecutive times. Participants wore one smartphone in a pocket that was attached to their belt and another one in their trouser pocket. All smartphones were equipped with a built-in accelerometer and the fall detection application. Four participants tested the application on a Samsung S3 and four tested the application on a Samsung S3 mini. Sensitivity scores were.75 (Samsung S3 belt),.88 (Samsung S3 mini trouser pocket), and.90 (Samsung S3 mini belt/Samsung S3 trouser pocket). Specificity scores were.87 (Samsung S3 trouser pocket),.91 (Samsung S3 mini trouser pocket),.97 (Samsung S3 belt), and.99 (Samsung S3 mini belt). These results suggest that an application on a smartphone can generate valid fall alarms when worn on a belt or in a trouser pocket. However, sensitivity should be improved before implementation of the application in practice.

Language: en